Assignment 5/MATH 247/Winter Due: Friday, February 19 in class (!) (answers will be posted right after class)

Similar documents
MATH 247/Winter Notes on the adjoint and on normal operators.

4 Inner Product Spaces

Chapter 9 Jordan Block Matrices

CHAPTER 4 RADICAL EXPRESSIONS

1 Onto functions and bijections Applications to Counting

Lecture 9: Tolerant Testing

18.413: Error Correcting Codes Lab March 2, Lecture 8

{ }{ ( )} (, ) = ( ) ( ) ( ) Chapter 14 Exercises in Sampling Theory. Exercise 1 (Simple random sampling): Solution:

Multiple Choice Test. Chapter Adequacy of Models for Regression

L5 Polynomial / Spline Curves

Summary of the lecture in Biostatistics

Ideal multigrades with trigonometric coefficients

Numerical Analysis Formulae Booklet

Chapter 4 Multiple Random Variables

UNIT 2 SOLUTION OF ALGEBRAIC AND TRANSCENDENTAL EQUATIONS

Laboratory I.10 It All Adds Up

Econometric Methods. Review of Estimation

α1 α2 Simplex and Rectangle Elements Multi-index Notation of polynomials of degree Definition: The set P k will be the set of all functions:

Functions of Random Variables

Chapter 3 Sampling For Proportions and Percentages

b. There appears to be a positive relationship between X and Y; that is, as X increases, so does Y.

Lecture 07: Poles and Zeros

hp calculators HP 30S Statistics Averages and Standard Deviations Average and Standard Deviation Practice Finding Averages and Standard Deviations

X ε ) = 0, or equivalently, lim

MATH 371 Homework assignment 1 August 29, 2013

A scalar t is an eigenvalue of A if and only if t satisfies the characteristic equation of A: det (A ti) =0

1 Mixed Quantum State. 2 Density Matrix. CS Density Matrices, von Neumann Entropy 3/7/07 Spring 2007 Lecture 13. ψ = α x x. ρ = p i ψ i ψ i.

Algorithms Theory, Solution for Assignment 2

ρ < 1 be five real numbers. The

Assignment 7/MATH 247/Winter, 2010 Due: Friday, March 19. Powers of a square matrix

X X X E[ ] E X E X. is the ()m n where the ( i,)th. j element is the mean of the ( i,)th., then

PTAS for Bin-Packing

QR Factorization and Singular Value Decomposition COS 323

Lecture 7. Confidence Intervals and Hypothesis Tests in the Simple CLR Model

STATISTICAL PROPERTIES OF LEAST SQUARES ESTIMATORS. x, where. = y - ˆ " 1

ENGI 4421 Joint Probability Distributions Page Joint Probability Distributions [Navidi sections 2.5 and 2.6; Devore sections

2SLS Estimates ECON In this case, begin with the assumption that E[ i

CLASS NOTES. for. PBAF 528: Quantitative Methods II SPRING Instructor: Jean Swanson. Daniel J. Evans School of Public Affairs

The internal structure of natural numbers, one method for the definition of large prime numbers, and a factorization test

PROJECTION PROBLEM FOR REGULAR POLYGONS

III-16 G. Brief Review of Grand Orthogonality Theorem and impact on Representations (Γ i ) l i = h n = number of irreducible representations.

12.2 Estimating Model parameters Assumptions: ox and y are related according to the simple linear regression model

Lecture Notes Types of economic variables

Simple Linear Regression

ESS Line Fitting

Chapter 5 Properties of a Random Sample

Applying the condition for equilibrium to this equilibrium, we get (1) n i i =, r G and 5 i

1 Lyapunov Stability Theory

MEASURES OF DISPERSION

Lecture 3 Probability review (cont d)

means the first term, a2 means the term, etc. Infinite Sequences: follow the same pattern forever.

CHAPTER VI Statistical Analysis of Experimental Data

We have already referred to a certain reaction, which takes place at high temperature after rich combustion.

UNIVERSITY OF OSLO DEPARTMENT OF ECONOMICS

n -dimensional vectors follow naturally from the one

Random Variables and Probability Distributions

MOLECULAR VIBRATIONS

ENGI 4421 Propagation of Error Page 8-01

MA 524 Homework 6 Solutions

Log1 Contest Round 2 Theta Complex Numbers. 4 points each. 5 points each

F. Inequalities. HKAL Pure Mathematics. 進佳數學團隊 Dr. Herbert Lam 林康榮博士. [Solution] Example Basic properties

CIS 800/002 The Algorithmic Foundations of Data Privacy October 13, Lecture 9. Database Update Algorithms: Multiplicative Weights

Midterm Exam 1, section 1 (Solution) Thursday, February hour, 15 minutes

Cubic Nonpolynomial Spline Approach to the Solution of a Second Order Two-Point Boundary Value Problem

5 Short Proofs of Simplified Stirling s Approximation

A New Method for Solving Fuzzy Linear. Programming by Solving Linear Programming

EECE 301 Signals & Systems

Lattices. Mathematical background

d dt d d dt dt Also recall that by Taylor series, / 2 (enables use of sin instead of cos-see p.27 of A&F) dsin

Third handout: On the Gini Index

ε. Therefore, the estimate

Multiple Regression. More than 2 variables! Grade on Final. Multiple Regression 11/21/2012. Exam 2 Grades. Exam 2 Re-grades

Q-analogue of a Linear Transformation Preserving Log-concavity

Derivation of 3-Point Block Method Formula for Solving First Order Stiff Ordinary Differential Equations

Multiple Linear Regression Analysis

7.0 Equality Contraints: Lagrange Multipliers

The number of observed cases The number of parameters. ith case of the dichotomous dependent variable. the ith case of the jth parameter

SPECIAL CONSIDERATIONS FOR VOLUMETRIC Z-TEST FOR PROPORTIONS

arxiv:math/ v1 [math.gm] 8 Dec 2005

Dr. Shalabh. Indian Institute of Technology Kanpur

Generalized Linear Regression with Regularization

The Arithmetic-Geometric mean inequality in an external formula. Yuki Seo. October 23, 2012

3D Geometry for Computer Graphics. Lesson 2: PCA & SVD

Investigation of Partially Conditional RP Model with Response Error. Ed Stanek

( ) 2 2. Multi-Layer Refraction Problem Rafael Espericueta, Bakersfield College, November, 2006

TESTS BASED ON MAXIMUM LIKELIHOOD

Qualifying Exam Statistical Theory Problem Solutions August 2005

Class 13,14 June 17, 19, 2015

Exercises for Square-Congruence Modulo n ver 11

Introduction to local (nonparametric) density estimation. methods

Multivariate Transformation of Variables and Maximum Likelihood Estimation

STA302/1001-Fall 2008 Midterm Test October 21, 2008

CS5620 Intro to Computer Graphics

THE ROYAL STATISTICAL SOCIETY 2016 EXAMINATIONS SOLUTIONS HIGHER CERTIFICATE MODULE 5

Some Different Perspectives on Linear Least Squares

Lecture 12 APPROXIMATION OF FIRST ORDER DERIVATIVES

Chapter 14 Logistic Regression Models

Discrete Mathematics and Probability Theory Fall 2016 Seshia and Walrand DIS 10b

A Study on Generalized Generalized Quasi hyperbolic Kac Moody algebra QHGGH of rank 10

STK4011 and STK9011 Autumn 2016

Transcription:

Assgmet 5/MATH 7/Wter 00 Due: Frday, February 9 class (!) (aswers wll be posted rght after class) As usual, there are peces of text, before the questos [], [], themselves. Recall: For the quadratc form x qx ( ) = qx (, x,..., x) = aj x xj ( X= ),, j= x we have the max-form qx ( ) = X AX ( A= ( a j ) ). y Moreover, f s a vertble max, ad X = Y ( Y = ), the y qx ( ) = qy ˆ( ) = qy ˆ(, y,..., y) = Y BY for the max B = A. We assume that ( ) A= a j s symmec: aj = aj for all, j =,..., ; as a cosequece, the B s also symmec. The quadratc form qx ( ) s ) postve defte, ) egatve defte, 3) defte f, respectvely: ) for all X 0, we have qx ( ) > 0 ; ) for all X 0, we have qx ( ) < 0; 3) there are X 0, X 0 such that qx ( ) > 0 ad qx ( ) < 0. These three qualtes are mutually exclusve; but a quadratc form may be oe of the above three kds. qx ( ) s ) postve semdefte, 5) egatve semdefte f, respectvely: ) for all X 0, we have qx ( ) 0; 5) for all X 0, we have qx ( ) 0. Our text says oegatve semdefte for postve semdefte. The termology s such that postve semdefte s more geeral tha postve defte ; ad smlarly for egatve semdefte. However, the mportat possblty s that qx ( ) may be postve semdefte, wthout beg postve defte. For stace,

qxy (, ) x xy y = + + s postve semdefte sce qxy (, ) = ( x+ y) 0 always; but 0 ad q( ) = q(, ) = ( ) = 0, thus q s ot postve defte. Of course, f defte. q s postve semdefte, the t caot be ether egatve defte, or (I the ma applcato aalyss (calculus), whe the quadratc form formed by the secod partals at a statoary pot s postve defte, oe has a local mmum; whe the quadratc form s egatve defte, oe has a local maxmum; whe t s defte, oe deftely has o local mmum, or a maxmum; ad whe t s semdefte but ot defte, oe does ot kow about local exemum wthout further vestgato. Ths wll be looked at later class.) Wth the quadratc form qx ( ) = X AX, we assocate the symmec blear form ( X, X ) defed by β ( X, X ) X A X β = β ( X, X ) = β ( X, X ) We have the detty qx) ( = β ( XX, ).. β s symmec the sese that [] Let qx ( ) = qx (, yzw,, ) = x y 3z + 5w+ 5xz+ xw+ wz (X Throughout ths questo, q deotes ths partcular quadratc form. x y = ). z w ) Wrte qx ( ) the symmec max form. r s ) Apply the coordate asformato X = Y, Y =, wth the t u asformg max = 0 5/ 7/37 0 0 0 0 0 8/37 0 0 0

ad wrte qx ( ) the form qxyzw (,,, ) = qrstu ˆ(,,, ) = byy j j (where we wrote y = r, y = s, y = t, y = u) wth sutable coeffcets. Use precse fracto 3 calculato, rather tha decmals; the result should come out precsely dagoal. b j, j= 3) Usg the max approprately, wrte each of the varables rstu,,, as a lear expresso of x, yzw,, ; ad verfy drectly the resultg detty qxyzw (,,, ) = qrs ˆ(,, tu, ). ) Usg the work doe so far, decde what kd q s the defteess classfcato (see above!). rove your asserto: show that q satsfes the defto gve above of the qualty (postve defte,, defte, ). Appealg to some theorem about that qualty s ot eough. 5) Cosder the symmec blear form β assocated wth q. Let X = 0, Y =. 0 3 Calculate β ( X, X)( = q( X)), β ( YY, ) ad β ( X, Y ), ad gve a formula for qax ( + by) the form of a homogeeous quadratc polyomal (quadratc form!) of the varables a ad b. The Gram-Schmdt orthogoalzato process Read secto 7.7, p. 6, the text. Whe you see er product space V, take V to be ay subspace of a stadard space, ad for the er product uv,, read dot product: X, Y = X Y. Note Remark o page 6 especally. Read Example 7.0 o the same page, but omt the ext example 7. (whch s a vector space of a kd that we have ot looked at yet). May also read solved problem 7., p.6. You wll eed to use Gram-Schmdt some problems that follow. 3

Recall: A max s called orthogoal f = = I ; equvaletly, f ; ad equvaletly, f the colums,..., of form a orthoormal system: = 0 whe j j; ad = =. If U = ( u,..., u ) ad V = ( v,..., v ) both m are orthoormal bases of a subspace V of some ( partcular whe V =, U = E, the stadard bass, ad V s aother orthoormal bass of ), the = [ U V ], the chage-of-bass max, s a orthogoal max. [] ) rove that f ad Q are orthogoal maces, the so are,, Q. ) Use Gram-Schmdt to fd a orthogoal max the frst colum of / / whch s =. / / Isuctos: Start wth ay bass B of the frst of whose vectors s. You may use some of the stadard vectors e,..., e as the last three vectors of B. Apply Gram- Schmdt to B. Clear deomators termedate results too. Use the symbol to / 3/ 3 3 deote a scaled vector: e.g., =. Scalg should be used also to 3/ 6 6 / 3 6 get smaller umbers: e.g.,. Normalze the system at the ed oly. Wrte 9 3 precse fractos ad square-roots whe ecessary; do ot use decmal approxmato. 3 3) Wth determed ), ad deotg the colums of by,,,, let U = spa(, ). Use ) (that s: do t start from scratch!) ad gve a bass of U, the orthogoal complemet of U.

7 ) For U as 3), ad Y =, determe proj ( 0 Y U ) ad proj U ( Y) (Ht: read Theorem 7.8, page 5 ( just before the secto o Gram-Schmdt), ad the Remark followg t. Take ote of the term Fourer coeffcet just before the remark. It wll come back later.) [3] We cosder the followg quadratc forms: q ( X ) = x + 3y z w + xy+ xz+ xw yz yw+ 6zw, q ( X ) = x + y + xy+ xz+ xw+ zw q ( X ) = 3 x 3z 3w yz yw + zw q ( X ) = 3x + 5y + 5z + 5w yz yw zw Treat each of the quadratc forms q = q, q,... as follows. ) Fd a orthogoal asformg max that turs qx ( ) to a puresquare (dagoal) form qy ˆ( ) terms of the ew varables Y ; wrte out what qy ˆ( ) s, ad what the relatoshp s betwee the old ad ew coordate vectors X ad Y. ) Determe whch kd q s the defteess classfcato. For each case whe the q s defte, or semdefte-o-defte, supply examples of vectors X that show the property questo. ( ) 3) Determe max qx X ( ) ad m qx X. ) Suppose that A s a symmec max, s a orthogoal max, ad A s a dagoal. Let B be a polyomal expresso of A : k B = bi 0 +ba +... + ba k. rove that B s symmec, ad B s dagoal. (Hts: f you fd t too hard, verfy the asserto frst for the specal cases B = A, A, I + A, I + 3A+ A,... utl the patter becomes clear.) 5

5) Cosder the quadratc form q ( X ) = 5 x + y + 3z + 3w + xy+ xz+ xw+ yz+ yw+ zw. Verfy that ( ) q5 X = X B X for the max B = A + A+ I, where A s the max for whch ( ) q X = X A X, wth q from above. Usg kowledge o q ( X), do the work asked for ), ) ad 3) for q ( X), wth a mmal amout of calculato oly. 5 6) Let A be a symmec max, qx ( ) = X AX the correspodg quadratc form. rove that the fucto QX ( ) defed by QX ( ) = qax ( X) s a quadratc form, ad ay orthogoal max that asforms q to pure-squares form, does the same for Q. Determe the values from the values λ for whch qx ( ) = λ y. = μ for whch qx ( ) = μ y ( X = Y ) = 6